What problem does it solve? LLM spend routed through the Caveman Cloud gateway lands in a single unlabeled bucket, making it impossible to see which jobs drive costs. This Skill inventories every LLM workflow in a repository, names each one with a valid slug, and wires workflow labels into the code so spend is grouped by workflow on the dashboard. ## Core Features & Use Cases - Workflow Inventory: Walks the repository from entry points (HTTP handlers, cron jobs, queue consumers, CLI scripts, eval harnesses, agent definitions) to find every LLM callsite. - Slug Naming & Proposal Table: Generates gateway-valid lowercase slugs that name the job rather than the technology, and presents a labeling table for user approval before any code changes. - Label Wiring: Adds the x-cave-workflow header or SDK workflow option at each callsite using the lightest mechanism available, covering Caveman SDKs, raw provider SDKs, wrapped agents, and raw HTTP. - Use Case: A team sees one large unlabeled LLM bill. Run this Skill to discover that spend actually comes from a support-reply handler, a nightly digest cron job, and an eval suite, then label each so the dashboard breaks costs down per workflow. ## Quick Start Ask the agent to discover and label all LLM workflows in this repository so Caveman Cloud groups spend by workflow.
